|Publication number||US7884964 B2|
|Application number||US 10/635,381|
|Publication date||Feb 8, 2011|
|Filing date||Aug 5, 2003|
|Priority date||Aug 5, 2003|
|Also published as||EP1505822A2, EP1505822A3, EP1505822B1, EP1505822B9, US20050030560|
|Publication number||10635381, 635381, US 7884964 B2, US 7884964B2, US-B2-7884964, US7884964 B2, US7884964B2|
|Inventors||Martin S. Maltz, Raja Bala, Lalit K. Mestha|
|Original Assignee||Xerox Corporation|
|Export Citation||BiBTeX, EndNote, RefMan|
|Patent Citations (15), Non-Patent Citations (3), Referenced by (3), Classifications (25), Legal Events (3)|
|External Links: USPTO, USPTO Assignment, Espacenet|
Embodiments generally relate to color rendering devices, such as printers and photocopy machines. Embodiments also relate to color rendering control and feedback methods and systems. Embodiments additionally relate to color adjustment methods and systems. Embodiments also relate to color sensors.
In many instances it can be necessary to produce faster, higher quality and more sophisticated rendering devices such as (e.g., printers, copiers, and the like). Additionally, customers demand the ability to reproduce their desired input images to achieve accurate output that does not vary over time (e.g., between pages or when output again at a later time) or among various rendering devices (e.g., among various printers on a network, between home and office marking devices, etc.). These considerations are more difficult to achieve with color rendering devices because of the greater number of variables involved and the sensitivity of these devices to environmental conditions, and also to issues concerning out-of-gamut colors.
In general, a gamut can be defined as the range of color that a rendering device (e.g., printer, copier, and the like) can produce, or the range of color that a color model can represent. In the color digital image processing arts, the space of colors that a printer (or other rendering device) is capable of reproducing is referred to as the “gamut” of the printer. Very often, colors specified in a digital image cannot be reproduced by a particular printer when it comes time to output the color digital image to the printer for purposes of rendering a hard copy printed output of the digital image.
The printer gamut (also referred to herein as the output gamut) is often limited when compared to the gamut of the source digital image (the input gamut) due to physical limitations of the printer's colorants. Colors in the input gamut but not in the output gamut are said to be “out-of-gamut” and must be accounted for before the digital image may be printed—i.e., areas in the output document where these colors are to be found may not simply be left blank in the final printed document.
Prior techniques for dealing with out-of-gamut colors include gamut clipping and gamut compression. In gamut clipping, all out-of-gamut colors are mapped to a color on the gamut “surface” while in-gamut colors are left unaltered. A common form of clipping involves a ray-based approach, wherein a ray is drawn from a desired out-of-gamut color to a point on a neutral axis. The location or point where the ray penetrates the gamut surface is the gamut-mapped color. Such a strategy preserves hue through the gamut-mapping operation. In gamut compression, both in-gamut and out-of-gamut colors are altered to map the entire range of image colors to the printer gamut. In practical systems, the gamut-mapping operation is often incorporated into a 3-dimensional lookup table for efficient computational processing.
A common problem in color printing is the stable rendition of colors over time. In particular, it is desirable when a color is situated near or on the boundary of a color gamut, to stabilize the resulting color rendering. Prior techniques for ensuring stable color control involve iterative closed-loop algorithms that apply a corrective feedback on 3-dimensional color signals. Such techniques can give rise to unstable or erroneous control for colors that are out-of-gamut, or on the gamut-surface.
Accordingly, it has been deemed desirable to provide a color rendering technique that overcomes the inability of traditional print engines and color rendering devices to control the printed color, particularly out-of-gamut colors given a particular gamut-mapping algorithm.
It is a feature of the present invention to provide improved color rendering methods and systems.
It is another feature of the present invention to provide improved control for out-of-gamut and index colors associated with color rendering devices and systems.
Aspects of the present invention related to methods and systems for controlling colors in an image processing device that are located external to a gamut are disclosed. A plurality of color values can be automatically provided as input to said image processing device, wherein said image processing device is under a control of a particular dimensional order, typically a three-dimensional order, but alternatively can be a two-dimensional order. Next, an operation can be performed dynamically determining which color value among the plurality of color values has attained a gamut limit. Thereafter, the particular dimensional order can be automatically reduced from, for example, a three-dimensional order to a two-dimensional or one-dimensional order based on determining which color value among said plurality of color values has attained said gamut limit, thereby providing improved control for colors that are located external to said gamut.
The plurality of color values analyzed using the present invention is generally associated with three colors—cyan, magenta, and yellow. Dynamic selection of appropriate dimensions, along with compensation/transformation inside an iterative controller results in improved out-of-gamut color control and ultimately, a more efficient and cleaner color rendering.
The accompanying figures, in which like reference numerals refer to identical or functionally-similar elements throughout the separate views and which are incorporated in and form part of the specification, further illustrate embodiments of the present invention.
The particular values and configurations discussed in these non-limiting examples can be varied and are cited merely to illustrate an embodiment of the present invention and are not intended to limit the scope of the invention.
In general, an iterative print and measure method can be implemented which determines an appropriate CMY combination for producing a specified L*a*b* (i.e., axes 102, 104 and 106). In the case of a four color printer, for example, the CMY can be processed by a suitable GCR/UCR strategy to obtain a corresponding CMYK. Ink limiting operations may also be performed. Such an iterative technique can, however, run into problems at the gamut surfaces, because the next trial CMY might be located outside the gamut of the printer.
Note that in general, the acronym “CMY” refers to cyan, magenta, and yellow, which are considered the subtractive primary colors, or a color space that describes colors in terms of their cyan, magenta, and yellow components. The subtractive primary colors of cyan, magenta and yellow are generally utilized to create reflective color. Cyan absorbs (i.e., subtracts) all red light, reflecting blue and green. Magenta, on the other hand, absorbs all green light, reflecting blue and red. Yellow absorbs blue light, reflecting red and green.
Because at the surface of the gamut, one of the three colorant coordinates (C, M, or Y) is fixed, the aforementioned three-dimensional problem can be converted into a two-dimensional problem. Consider, for example, the case of an out of gamut red. Producing such a color would require a negative amount of cyan. Because achieving a negative amount of cyan is physically impossible, such a color is out of reach of the printer's gamut. To obtain the gamut mapped color, it is necessary to set C=0, and vary M and Y to obtain the “best” gamut mapped red.
There are several possible strategies for gamut mapping. Ray-based methods represent one possible gamut mapping strategy. In ray-based approaches, a ray can be drawn from a desired color to a point on a neutral axis, which results in the preservation of hue. The location or point where the ray penetrates the gamut surface is therefore the desired color.
To obtain the CMY combination that produces this color, it is necessary to set C=0 and vary M and Y to get the desired color. This color can be found by utilizing an NCD coordinate system, where N is a vector coincident with the desired ray, and C and D are two orthogonal unit vectors at right angles to N. A suitable transformation of the form L*a*b*=M*NCD+V can always be determined in order to transform between NCD and L*a*b* vectors, where M is a 3×3 matrix, and V is a vector. The inverse of this transform can also be found. This transformation can be designated by a symbol, T. In the NCD coordinate system, the desired color results where the C and D component of the NCD vector is equivalent to zero.
Next, a two-dimensional Jacobian matrix can be determined, which describes the printer response in terms of the derivatives of the C and D components of the NCD vector with respect to M and Y at the point where the ray penetrates the surface of the gamut. An initial estimate for the M and Y can also be obtained to produce the desired color. Such estimates can be performed utilizing a variety of techniques. For example, a convenient estimation technique involves the use of an approximate printer model. Because such a model is only an approximate model, printing an initial CMY combination generally produces a color (e.g., E1) that does not lie at the desired point on the surface of the printer gamut.
Utilizing the two-dimensional Jacobian matrix, a gain matrix can be found which is utilized to calculate an improved estimate of M and Y, given the values of the C and D components of the NCD vector for the printed patch. Then, utilizing an iterative print and measure process (e.g., E2 and E3), values of M and Y can be determined, which make the C and D components of the NCD vector equal to zero. This is the desired color.
Another gamut mapping approach involves searching for the color on the surface of the gamut that is nearest to the specified color. This case is usually handled as discussed above, except that the ray (e.g., ray 108) and N vector are selected to be perpendicular to the gamut surface. If the gamut mapped color lies on a gamut edge or vertex, however, a modified approach is used. If the gamut mapped color lies on a gamut vertex, all the CMY components are known, and nothing further need be done. If it lies on an edge of the gamut, the problem becomes one dimensional. In this case, the ray and N vector go from the edge to the specified color, and are perpendicular to the edge. The C unit vector can be located parallel to the edge. Only one colorant is varied, such that the derivative required be evaluated is the derivative of the C component of the NCD vector with respect to the colorant being varied.
Still, a further gamut mapping technique involves searching for the color on the surface of the gamut which is nearest to the specified color, but is constrained to be of the same hue. This case can be handled similar to the previously discussed cases, except that the ray and N vector would lie in a plane of the desired hue, and would be located perpendicular to the contour formed by the intersection between this plane and the gamut surface.
The aforementioned embodiments all involve a dedicated gamut mapping function that determines the surface point, as well as the axes of the NCD coordinate system. In some applications, it may not be feasible or convenient to invoke such a function. In an alternative embodiment, a standard International Color Consortium (ICC) framework can be utilized to realize an embodiment of the present invention. A specified input color can be mapped through a BtoA tag in one of the rendering intents (e.g., preferably a calorimetric intent) of the ICC profile to obtain an initial CMY estimate.
Mapping the CMY point through the AtoB tag produces a corresponding color on the gamut surface in L*a*b* space. Only the NCD axes remain to be determined. Ideally such axes would be calculated for every new input point. In practice, however, it is probably sufficient to store one set of axes for each gamut face and edge. These can be computed and stored a priori in a private ICC tag using information from gamuts that are representative of the device under consideration.
If the specified color is well within the gamut, a standard 3-D control approach can be applied. For example, a 3×3 gain matrix can be derived to iteratively estimate the CMY combination that produces the desired L*a*b* color.
As the desired color approaches the gamut boundary, a proper transition should be ensured from the 3D strategy to the 2-D (or 1-D) strategy described herein with respect to embodiments.
One approach would be to apply a linear interpolation between the outputs of the 3-D and 2-D iterations, where the interpolation weight depends on the distance to the gamut surface. The latter can be estimated by examining the minimum and maximum values of the initial CMY estimate, and their deviation from 0 and 100% area coverage, respectively. Alternatively, the distance to the gamut surface in L*a*b* can be utilized. A measure of the stability of the 3-D iterative algorithm could also be utilized as an indicator of how close the point is to the gamut surface.
System 200 includes a transformation module 204 (i.e., T) which receives input in the form of desired L*a*b* values of memory color, as indicated at block 202. Output 206 from transformation module 204 is generally in the form of desired NCD values of the memory color, which in turn are input to an adder 208. Output from adder 208 is input to an iterative controller 210 whose output 212 reflects compensated CMY values that require correction for printer/media variations.
Output 212 is generally provided to a printer 214 or another graphical or textual rendering device (e.g., a photocopy machine). Output 216 is then generated by printer 214 in the form of patches, which are provided as input to a color sensor 218. Output 220 from color sensor 218 can then generate L* a* b* values of measured patches, as indicated by block 220. Output 220 can then be provided as input to a transformation module 224 (i.e., also T) whose output provides measured NCD values of the memory color. Output from transformation module 224 is then fed as input to adder 208, thereby completing a feedback control loop. Note that system 200 can be utilized not only with online sensors (e.g., color sensor 218), but may also be utilized with offline sensors. utilized with offline sensors.
Transformation modules 204 and 224 generally implement a transformation function, T. An example of a transformation function, T, can be illustrated as follows: Consider a very bright red, where L*a*b*=[100, 50, 50], gamut mapped along a ray that goes to L*a*b*=[50, 0, 0]. For this case, the following matrices M and V can be designated via equations (1) and (2) below:
Variables N, C, and D are governed by a number of parameters. For example, in this transformation example, N extends from L*a*b*=[50,0,0] to L*a*b*=[100,50,50] so that deltaL*a*b* is proportional to [1,1,1]. C is located at right angles in a plane of constant L* so that deltaL*a*b*=[0,−1,1]. Finally, D is N X C, which becomes [2,−1,−1] (i.e., X is the vector cross product).
Refer also to N, C, and D of
The process depicted in
Thereafter, as illustrated at block 316, a Jacobian matrix can be determined at the nominal CMY values utilizing the printer model described earlier with respect to block 312. This Jacobian matrix involves the derivatives of C and D in the NCD orthogonal space with respect to the appropriate axes in CMY space. For example, in the cyan=0 example considered above, the derivatives would be taken with respect to the magenta and yellow axes.
Performing this calculation requires the specification of the T transformation, which can be obtained as indicated at blocks 306 and 308. As depicted at block 306, an operation is performed in which the NCD orthogonal vector space is searched for the desired memory color and desired gamut mapping strategy.
Thereafter, as indicated at block 308, a transformation T can be determined by inverting a biased L*a*b* to an NCD transform. An example of the operation depicted at block 308 is mathematically provided by equation (4) indicated above. Next, as indicated at block 318, the gain matrix can be determined utilizing the Jacobian matrix calculated earlier. This also requires the T matrix calculated as depicted at block 308.
The gain matrix information via the operation indicated at block 318 and the nominal CMY values as depicted at block 314 are next fed to block 328 on
The operation described at block 328 generally involves processing the error vector calculated via the operation indicated at block 310 in a feedback controller that utilizes the gain matrix and nominal CMY values to obtain new CMY values. In the first pass through the iteration loop, however, no error has yet been calculated. In this case, block 328 just passes the nominal CMY values on to block 330, which produces and prints an image containing test patches. Next, as depicted at block 332, L*a*b* values from the test patches can be measured using, for example, a spectrophotometer.
As indicated at block 334, these L*a*b* vectors can be transformed to NCD coordinates utilizing the T transformation matrix, obtained from block 308 through continuation block 320 (“C”). The L*a*b* values obtained the operation illustrated at block 332, and the CMY values as indicated at block 328 can also be stored, as depicted at block 336. When the iterative process has reduced the error to a satisfactory level (the output of comparison block 322 is “yes”), block 322 sends a signal to block 336 through continuation block 324 (“B”) to “write out” the CMY and L*a*b* data in a suitable format. Customer images can also be appropriately modified at this time. The process then terminates, as indicated at block 338.
If the process does not terminate, the new NCD values are sent to block 310 through continuation block 320 (i.e. “D”). Block 310 compares the C & D values with the desired C & D values and calculates an error vector. The desired C and D values are usually 0, but more generally could be a function of the NCD orthogonal vector space determined in block 306.
As indicated in block 322, the error calculated in block 310 can be compared to the limit. If the error is above the limit, the error vector can be fed to block 328 through continuation block 326 (“A”), where it can be processed in the feedback controller. An example of a feedback controller that may be utilized in this manner is the iterative controller 210 described earlier with respect to
In an alternative embodiment of the present invention, the operations depicted at blocks 312 and 314 can be processed in association with the operation depicted at block 328. In other words, the operations of block 312 and 314 can be processed within a feedback controller. For such implementations, the gain matrix for each memory color generally contains 3×3 elements. The selection of appropriate elements can be established within the feedback controller, dependent upon which CMY values reach the boundary.
It can be appreciated from the foregoing that embodiments of the present invention can be implemented in the context of modules. In the computer programming arts, a module can be typically implemented as a collection of routines and data structures that performs particular tasks or implements a particular abstract data type.
Modules generally are composed of two parts. First, a software module may list the constants, data types, variable, routines and the like that can be accessed by other modules or routines. Second, a software module can be configured as an implementation, which can be private (i.e., accessible perhaps only to the module), and that contains the source code that actually implements the routines or subroutines upon which the module is based. Thus, for example, the term module, as utilized herein generally refers to software modules or implementations thereof. Such modules can be utilized separately or together to form a program product that can be implemented through signal-bearing media, including transmission media and recordable media.
The logical operations depicted in
Graph 500 thus generally illustrates convergence plots, wherein cyan and yellow are limited to 0 and 100% respectively. In graph 500, only magenta is variable. A final steady state value for magenta, according to graph 5900, is approximately, 86.7%. Thus the gamut mapped CMY values in percent are [0 86.7 100]. Corresponding gamut mapped L*a*b* values are given by [60.33 55.85 55.85] as indicated in equation (5) below:
Based on the foregoing it can therefore be appreciated that when the consistency of desired memory color is required, an appropriate user interface can be utilized to produce desired L*a*b* values for a memory color control system. Because the memory color control system possesses an inverse characterization table (e.g., L*a*b* to CMY/CMYK), a correct strategy for gamut mapping in association with the desired L*a*b* values can be utilized to obtain three colorant CMY (i.e., cyan, magenta, yellow) coordinates.
If a patch with CMY values is printed and the color measured utilizing a spectrophotometer, the resulting measured L*a*b* value may not correspond to the desired L*a*b* value. Such a discrepancy may result from two possible reasons: (1) gamut mapping of colors to the gamut surface, and/or (2) variations in the engine caused by associated media and processes. Regarding reason (1), gamut mapping of colors to the gamut surface is necessitated primarily by device physics, and can be accomplished by a number of techniques known in the art. After the gamut-mapping step, a comparison of measured colors is generally performed with gamut-mapped L*a*b* values within a memory color control module.
To address reason (2), stable control of the gamut-mapped colors with respect to engine variations is accomplished via the aforementioned order reduction, which converts a three-dimensional control problem to a two- or a one-dimensional control problem. This utilizes a separate algorithm to produce the correct amount of CMY values so that when the patches are printed with those particular CMY values and measured with a spectrophotometer, colors can correspond to the desired L*a*b* after gamut-mapping.
The order reduction can be accomplished by selecting and eliminating appropriate colors at the gamut surface. For example, consider an out of gamut red. Generating such a color would require a negative amount of cyan. Because this action is physically impossible, such a color is out of the printer's gamut. To find the gamut mapped color, it may be necessary to set C=0, and vary M and Y to obtain the “best” gamut mapped red. Varying M and Y can be accomplished within an iterative controller as described herein.
It is appreciated that a variety of spectrophotometers can be implemented in accordance with embodiments of the present invention. One example of such a spectrophotometer is disclosed in U.S. Pat. No. 6,567,170, “Simultaneous Plural Colors Analysis Spectrophotometer,” which issued to Tandon, et al on May 20, 2003. U.S. Pat. No. 6,567,170 is assigned to Xerox Corporation and is incorporated herein by reference.
It will be appreciated that various of the above-disclosed and other features and functions, or alternatives thereof, may be desirably combined into many other different systems or applications. Also that various presently unforeseen or unanticipated alternatives, modifications, variations or improvements therein may be subsequently made by those skilled in the art which are also intended to be encompassed by the following claims.
|Cited Patent||Filing date||Publication date||Applicant||Title|
|US5704026||Jan 13, 1997||Dec 30, 1997||Eastman Kodak Company||Method and apparatus for determining a gamut boundary and a gamut descriptor|
|US5832109 *||Sep 3, 1996||Nov 3, 1998||Agfa Gevaert N.V.||Method and apparatus for calculating color gamuts|
|US5872898||Aug 30, 1996||Feb 16, 1999||Agfa Gevaert N.V.||Method and apparatus for calculating color gamuts|
|US6229916 *||Sep 30, 1998||May 8, 2001||Fuji Photo Film Co., Ltd.||Color transformation look-up table|
|US6236474 *||May 22, 1998||May 22, 2001||Xerox Corporation||Device independent color controller and method|
|US6335983||Sep 28, 1998||Jan 1, 2002||Eastman Kodak Company||Representing an extended color gamut digital image in a limited color gamut color space|
|US6342951||Dec 16, 1998||Jan 29, 2002||Xerox Corporation||Gamut mapping algorithm using inverted gamma function|
|US6400843||Apr 22, 1999||Jun 4, 2002||Seiko Epson Corporation||Color image reproduction with accurate inside-gamut colors and enhanced outside-gamut colors|
|US6437792 *||Jan 20, 2000||Aug 20, 2002||Sony Corporation||Image processing apparatus and method, color gamut conversion table creating apparatus and method, storage medium having image processing program recorded therein, and storage medium having recorded therein color gamut conversion table creating program|
|US6560356||Apr 17, 1998||May 6, 2003||Mitsubishi Denki Kabushiki Kaisha||Color gamut compression method and color gamut compression device|
|US6567170||Jun 25, 2001||May 20, 2003||Xerox Corporation||Simultaneous plural colors analysis spectrophotometer|
|US6618499||Jun 1, 1999||Sep 9, 2003||Canon Kabushiki Kaisha||Iterative gamut mapping|
|US6750992 *||Jan 12, 1999||Jun 15, 2004||Richard A. Holub||System for distributing and controlling color reproduction at multiple sites|
|US7167277 *||Jun 25, 2003||Jan 23, 2007||Fujitsu Limited||Color data conversion method, color data conversion apparatus, storage medium, device driver and color conversion table|
|US20040096104 *||Jul 28, 2003||May 20, 2004||Samsung Electronics Co.., Ltd.||Method of color correction|
|1||European Partial Search Report for EP Application 04018583.7 dated Oct. 30, 2009.|
|2||Inui, M., "Fast Algorithm for Computing Color Gamuts," Color Research and Application (1993) October, 18(5):341-348.|
|3||Morovic, J. et al., "The Fundamentals of Gamut Mapping: A Survey," Journal of Imaging Science and Technology (2001) 45(3):283-290.|
|Citing Patent||Filing date||Publication date||Applicant||Title|
|US8760060 *||Nov 13, 2009||Jun 24, 2014||Prism Projection, Inc.||Solid state light fixture with enhanced thermal cooling and color mixing|
|US8870319||May 2, 2012||Oct 28, 2014||Xerox Corporation||System and method for printing with ink limiting|
|US20110012512 *||Nov 13, 2009||Jan 20, 2011||Garrett J. Young||Solid state light fixture with enhanced thermal cooling and color mixing|
|U.S. Classification||358/1.9, 358/518, 345/593, 358/521, 345/590, 345/690, 382/164, 358/530, 382/167, 345/549, 382/162, 358/515, 345/591, 358/516, 358/448, 358/1.13, 382/254, 358/519, 382/232, 358/525|
|International Classification||H04N1/60, G06F15/00, G06K9/00|
|Aug 5, 2003||AS||Assignment|
Owner name: XEROX CORPORATION, CONNECTICUT
Free format text: ASSIGNMENT OF ASSIGNORS INTEREST;ASSIGNORS:MALTZ, MARTIN S.;BALA, RAJA;MESTHA, LALIT K.;REEL/FRAME:014383/0220
Effective date: 20030801
|Aug 31, 2004||AS||Assignment|
Owner name: JPMORGAN CHASE BANK, AS COLLATERAL AGENT, TEXAS
Free format text: SECURITY AGREEMENT;ASSIGNOR:XEROX CORPORATION;REEL/FRAME:015722/0119
Effective date: 20030625
Owner name: JPMORGAN CHASE BANK, AS COLLATERAL AGENT,TEXAS
Free format text: SECURITY AGREEMENT;ASSIGNOR:XEROX CORPORATION;REEL/FRAME:015722/0119
Effective date: 20030625
|Jul 18, 2014||FPAY||Fee payment|
Year of fee payment: 4